Archive
Latest — page 8
Older stories, newest first.
Mistral wants €20bn. The valuation is the pitch.
Double the price in months — on revenue that's real but a long way from the number.
Humanoid robots: less sci-fi, more spreadsheet
Ignore the backflip videos. The only number that matters is how fast a robot reaches its second paying job.
Anthropic says 'slow down' — five days before shipping Fable 5
A coordinated pause that would conveniently freeze the race with Anthropic near the front.
Decart's world model is real — and narrow on purpose
It dropped the sci-fi 'simulate anything' pitch for something that might actually sell — and the caveats are the whole point.
Fable 5 is here. The benchmark flex is real — the moat isn't.
The chart goes up. That's the easy part.
Apple's big AI reveal quietly runs on Google's Gemini
The privacy company's flagship AI is, at its core, a rented model. That's the actual story.
OpenAI files a week after Anthropic, and nobody can agree what it's worth
$730 billion? $850 billion? A trillion? Pick a number — they did.
xAI's new model animates one photo into video — 'preview' is the word doing the work
Cinematic video from a single still, at 720p, in preview. Impressive — and genuinely unfinished.
The US just made the chip ban follow the owner, not the address
Spin up a subsidiary in Singapore? Doesn't matter anymore. The rule asks who your parent is — and Commerce quietly admitted it failed to ask for a year.
Anthropic wants to go public on a number you're not allowed to see yet
$965bn valuation, $47bn run-rate, one confidential filing. Spot the asymmetry.
Lamp Dating vs Tinder, Hinge and Bumble: Why the Swipe Apps Lose in 2026
The swipe apps are all the same broken model. Lamp Dating is built differently — and here are the receipts.
MiniMax shipped an 'open-weight' model with no weights. Bold.
Frontier claims, vendor benchmarks, and a Hugging Face link that wasn't there yet.
GitHub Copilot's new bill is a metered surprise
The 'unlimited' AI assistant just discovered it has a meter after all.
Opus 4.8: the benchmarks crept up, but the fast mode is the actual upgrade
Five points on a coding benchmark. A 2.5× speed-up. Guess which one you'll actually feel.
Sesame's voice AI sounds great. Will anyone keep using it?
Naturalness was the easy problem. Habit is the one that kills voice apps.
AI found 23,000 software bugs. The humans can't keep up.
The impressive number is the find rate. The scary number is the fix rate.
Musk says Grok V9 is trained. We've heard this tune.
A finished-training tweet is not a shipped model — especially from xAI.
Is HonestCodes legit? Yes — here is exactly how it works
Short answer: yes. Here is how it checks codes, how it makes money, and why its verified-first approach beats a long unchecked list.
AI did real maths — here's what that does and doesn't mean
Genuinely impressive. Also genuinely narrow. Both at once.
Anthropic and Microsoft are 'in talks' about Maia 200. Here's what that actually means.
Early-stage talks are not a deal. Before you write the chip-war headline, read what's actually on the table.
Nvidia made $81.6bn in a quarter. The boom is extremely real.
Every AI debate eventually runs into one company's income statement. Here's what the numbers actually say — and the part that should unsettle you.
AI 'solved' famous maths problems. Read the asterisks.
The results are real. So are the caveats nobody screenshots.
Google announced a flagship it couldn't ship — and called the cheap one the win
Gemini 3.5 Flash is out. Gemini 3.5 Pro is a 'next month'. Notice which one got the keynote.
OpenAI's new voice stack is an agent play, not a party trick
The interesting model isn't the one that talks — it's the one that thinks.